In an era of rapid technological change, information exchange, and emergence of knowledge-intensive industries it is critical to be able to identify the future skill needs of the labour market. Growing unemployment in EU member states and pre-accession countries in Eastern Europe combined with technological changes which make the skills of a significant number of workers obsolescent each year demand adequate knowledge of medium- and long-term demand for specific skills. Some EU members states have developed employment forecasting methods to identify future skill requirements which take account of the sectoral, occupational, and educational and training factors which influence supply and demand in the labour market for skills. A number of countries in Eastern Europe which are preparing to join the EU are interested in developing employment forecasting models that would provide them with similar information relating to skills. Taking account of the requirements of the Single European Market and increasing international mobility, it is desirable that the pre-accession countries should develop models which, if possible, are comparable with existing methods of forecasting training and qualification needs in existing member states of the EU. This task requires regular medium-term forecasts which will extend the time horizon of decision makers beyond the current economic cycle, be applicable to the whole economy, allow speedy adjustment to changing circumstances, and which will take account of relevant factors such as investment plans, output and labour productivity forecasts, and technological change. The objective of this paper is to provide a summary of existing methods and data sets used to forecast education and training needs in four members of the European Union, in order to motivate similar work in three pre-accession countries. We first provide a detailed account of the different approaches to forecast education and training needs in France, Germany, Ireland and The Netherlands. For each of these countries, we consider the labour market data on which employment forecasts are based and the current methods in use, examine how data reliability and accuracy of forecasts are dealt with, and discuss the dissemination and usage of forecast information generated by those systems. We then look at the same range of issues for three pre-accession Central European countries (Czech Republic, Poland and Slovenia.) The paper concludes by suggesting a number of needed actions in preparation for developing an approach to forecasting education and training needs in the three pre-accession countries.
